日本地球惑星科学連合2018年大会

講演情報

[EE] Eveningポスター発表

セッション記号 A (大気水圏科学) » A-AS 大気科学・気象学・大気環境

[A-AS05] Precipitation Extreme

2018年5月23日(水) 17:15 〜 18:30 ポスター会場 (幕張メッセ国際展示場 7ホール)

コンビーナ:谷田貝 亜紀代(弘前大学大学院理工学研究科)

[AAS05-P03] SPATIAL CORRELATION STRUCTURE OF APHRODITE IN DATA SCARCE REGIONS

*Raazia Attique1 (1.INTEGRATION, Umwelt and Energie, Country Office, Pakistan)

Lack of observed gauged data in several parts of India and Pakistan is a key limitation to APHRODITE’s (Asian Precipitation - Highly-Resolved Observational Data Integration Towards Evaluation) daily gridded precipitation dataset. Scarce gauge station data raises a question on the reliability of gridded precipitation product to be used in further research. For this reason, the quality and reliability of the observed gauged as well as the APHRODITE (version V1101 available at 0.25o resolution) was assessed over the data scarce southern coastal region, ‘Gwadar-Ormara’ basin in Pakistan using cross- correlogram. An individual spatial correlogram (correlation coefficient vs. distance) was produced for both in-situ and APHRODITE data to check which dataset shows high correlations for closely located points and smaller values for points at far distances, thus following the Tobler’s first law of Geography (Tobler, 1970). To produce spatial correlograms, daily rainfall values in mm/day from 6 stations namely Pasni, Shadikaur, Tank, Hore, Chibkalamati and Basolmasjid were used for the years 1988-1991. The four-year period was chosen since only in this period daily time series were available without gaps for maximum gauge stations in the study area. Results of the spatial correlation of the two dataset showed that in-situ follows Tobler’s law better than the APHRODITE. The correlogram of in-situ showed structured shape, i.e., with the increase in the distance the correlation coefficient decreases. On the other hand, APHRODITE showed a linear highly correlated graph. Based on this result, it was concluded that APHRODITE can show results far – off from the ground reality due to its processed dataset and may provide unrealistic spatial patterns of precipitation in regions especially with scarce coverage of climate stations compared to the real observed dataset.


Ref: Tobler W. (1970) "A computer movie simulating urban growth in the Detroit region". Economic Geography, 46(Supplement): 234-240.